Using Relaxation to Fuse RFID and Vision for Object Tracking Outdoors


Fusion of Radio Frequency Identification (RFID) with Computer Vision (CV) can significantly improve performance in applications of autonomous vision and navigation, activity analysis, site monitoring, and especially in outdoor environments. RFID and CV provide both overlapping and unique information for deciding on object identity, location, and motion. We use relaxation to control the integration of information from CV, RFID, and naïve physics. Work site analysis must proceed even when information from one sensor or information source is unavailable at some time instances. Simulations show how fusion can greatly increase tracking performance while also reducing computational cost. Test cases show how fusion can solve some difficult tracking problems outdoors.

Keywords: Tracking, stereo, RFID, fusion, relaxation, site monitoring.